CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ke...@cvs.openbsd.org 2009/05/12 00:57:07 Modified files: net/lftp : Makefile distinfo Removed files: net/lftp/patches: patch-src_Http_cc Log message: update to 3.7.13. tested on sparc64 by pea@, thanks!
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2009/05/12 06:18:04 Modified files: net/quagga : Makefile distinfo net/quagga/patches: patch-configure patch-doc_Makefile_in patch-ospfclient_Makefile_in Removed files: net/quagga/patches: patch-bgpd_bgp_aspath_c patch-bgpd_bgp_attr_c patch-bgpd_bgp_community_c patch-bgpd_bgp_debug_c patch-bgpd_bgp_ecommunity_c patch-bgpd_bgp_mplsvpn_c patch-bgpd_bgp_packet_c patch-bgpd_bgp_route_c patch-bgpd_bgp_vty_c patch-bgpd_bgpd_c Log message: bugfix update to 0.99.12; rolls in patches for the problem where 4-byte ASNs with long prepends are seen.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ja...@cvs.openbsd.org 2009/05/12 13:51:42 Modified files: mail/postfix/snapshot: Makefile distinfo Log message: update snapshot; from brad
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ja...@cvs.openbsd.org 2009/05/12 14:03:08 Modified files: mail/postfix/stable: Makefile distinfo mail/postfix/stable/files: postfix-install mail/postfix/stable/patches: patch-master.cf mail/postfix/stable/pkg: PLIST Log message: upgrade to postfix 2.6.0; from brad
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2009/05/12 15:41:16 Modified files: devel : Makefile Log message: +libvstr
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2009/05/12 15:38:59 Log message: import libvstr: Vstr is a string library designed for network communication. Its design uses blocks of ptr+length data, so adding, substituting, and deleting data are all fast operations. It has a full API of all the usual string tasks: searching, comparing, splitting, substitution, converting between upper and lower case, and parsing numbers and strings. Has a full POSIX and ISO 9899:1999 compliant printf() implementation including gcc warning compatible custom formatters. Status: Vendor Tag: sthen Release Tags: sthen_20090512 N ports/devel/libvstr/Makefile N ports/devel/libvstr/distinfo N ports/devel/libvstr/pkg/PFRAG.shared N ports/devel/libvstr/pkg/DESCR N ports/devel/libvstr/pkg/PLIST No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:20:52 Modified files: databases/sqlite3: Makefile databases/sqlite3/patches: patch-src_os_unix_c Log message: enable load-extension for ports starting to require it tested by bernd@ in a bulk build, thanks!
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:26:26 Modified files: databases/p5-DBD-SQLite: Makefile distinfo databases/p5-DBD-SQLite/patches: patch-dbdimp_c Added files: databases/p5-DBD-SQLite/patches: patch-Makefile_PL Log message: update to version 1.25 tested by bernd@ in a bulk build, thanks! early test and late reminder by landry@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:33:06 Log message: import p5-Hash-Util-FieldHash-Compat Hash::Util::FieldHash using perltie. When Hash::Util::FieldHash is available it will use that instead. This way code requiring field hashes can benefit from fast, robust field hashes on Perl 5.10 and newer, but still run on older perls that don't ship with that module. Status: Vendor Tag: simon Release Tags: simon_20090513 N ports/devel/p5-Hash-Util-FieldHash-Compat/Makefile N ports/devel/p5-Hash-Util-FieldHash-Compat/distinfo N ports/devel/p5-Hash-Util-FieldHash-Compat/pkg/PLIST N ports/devel/p5-Hash-Util-FieldHash-Compat/pkg/DESCR No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:33:20 Modified files: devel : Makefile Log message: add p5-Hash-Util-FieldHash-Compat and p5-MooseX-Clone
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:33:14 Log message: import p5-MooseX-Clone Out of the box Moose only provides very barebones cloning support in order to maximize flexibility. This role provides a clone method that makes use of the low level cloning support already in Moose and adds selective deep cloning based on introspection on top of that. Attributes with the Clone trait will handle cloning of data within the object, typically delegating to the attribute value's own clone method. Status: Vendor Tag: simon Release Tags: simon_20090513 N ports/devel/p5-MooseX-Clone/Makefile N ports/devel/p5-MooseX-Clone/distinfo N ports/devel/p5-MooseX-Clone/pkg/PLIST N ports/devel/p5-MooseX-Clone/pkg/DESCR No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:44:40 Log message: import p5-MojoX-Session MojoX::Session is a session management for Mojo. Based on CGI::Session and HTTP::Session but without any dependencies except the core ones. Status: Vendor Tag: simon Release Tags: simon_20090513 N ports/www/p5-MojoX-Session/Makefile N ports/www/p5-MojoX-Session/distinfo N ports/www/p5-MojoX-Session/pkg/PLIST N ports/www/p5-MojoX-Session/pkg/DESCR No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:49:27 Log message: import p5-MojoX-Renderer-TT This module implements a Template Toolkit based renderer for Mojo. Status: Vendor Tag: simon Release Tags: simon_20090513 N ports/textproc/p5-MojoX-Renderer-TT/Makefile N ports/textproc/p5-MojoX-Renderer-TT/distinfo N ports/textproc/p5-MojoX-Renderer-TT/pkg/PLIST N ports/textproc/p5-MojoX-Renderer-TT/pkg/DESCR No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:49:33 Modified files: textproc : Makefile Log message: add p5-MojoX-Renderer-TT
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:57:57 Modified files: math : Makefile Log message: add p5-Math-Complex
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 17:57:53 Log message: import p5-Math-Complex This package lets you create and manipulate complex numbers. By default, Perl limits itself to real numbers, but an extra use statement brings full complex support, along with a full set of mathematical functions typically associated with and/or extended to complex numbers. Status: Vendor Tag: simon Release Tags: simon_20090513 N ports/math/p5-Math-Complex/Makefile N ports/math/p5-Math-Complex/distinfo N ports/math/p5-Math-Complex/pkg/PLIST N ports/math/p5-Math-Complex/pkg/DESCR N ports/math/p5-Math-Complex/patches/patch-Makefile_PL No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 18:07:52 Log message: import p5-Bot-BasicBot Basic bot system designed to make it easy to do simple bots, optionally forking longer processes (like searches) concurrently in the background. Status: Vendor Tag: simon Release Tags: simon_20090513 N ports/net/p5-Bot-BasicBot/Makefile N ports/net/p5-Bot-BasicBot/distinfo N ports/net/p5-Bot-BasicBot/patches/patch-t_connect_t N ports/net/p5-Bot-BasicBot/pkg/PLIST N ports/net/p5-Bot-BasicBot/pkg/DESCR No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 18:08:01 Modified files: net: Makefile Log message: add p5-Bot-BasicBot
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 18:14:55 Modified files: misc : Makefile Log message: add p5-File-MMagic-XS
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: si...@cvs.openbsd.org 2009/05/12 18:14:51 Log message: import p5-File-MMagic-XS This is a port of Apache2 mod_mime_magic.c in Perl, written in XS with the aim of being efficient and fast especially for applications that need to be run for an extended amount of time. Status: Vendor Tag: simon Release Tags: simon_20090513 N ports/misc/p5-File-MMagic-XS/Makefile N ports/misc/p5-File-MMagic-XS/distinfo N ports/misc/p5-File-MMagic-XS/pkg/DESCR N ports/misc/p5-File-MMagic-XS/pkg/PLIST No conflicts created by this import
Re: boost
On 2009-05-09, Markus Hennecke markus-henne...@markus-hennecke.de wrote: Andrej Elizarov wrote: Lucky day, I need boost for wt. It seems wt depend on boost at least v1.35. Markus, do you need any help with porting wt? Sure, I could use some help. If you use the asio port, then boost 1.34.1 will be enough I think. I Was this a separate port just to avoid touching the main boost port, or is there another reason? Thanks Stuart
pkg_info question
hi there, for any given package pkg_info starts its info like this, e.g.: amaaq$ pkg_info iodbc-3.52.4p2 Information for inst:iodbc-3.52.4p2 snip my question is, what does that inst: prefix mean, and can it be something else in different situations? -f -- to refuse to decide is a decision.
Re: pkg_info question
On 2009/05/12 15:07, frantisek holop wrote: hi there, for any given package pkg_info starts its info like this, e.g.: amaaq$ pkg_info iodbc-3.52.4p2 Information for inst:iodbc-3.52.4p2 snip my question is, what does that inst: prefix mean, and can it be something else in different situations? -f -- to refuse to decide is a decision. installed on local system. you can also have a remote URL there.
Re: pkg_info question
For locally installed packages the prefix is inst:. If you invoke pkg_info with the name of a pkg that's not installed, it will be prefixed with the url of the PKG_PATH mirror it got the pkg from... i.e. I don't have the mod_auth_ldap package installed, so... % pkg_info mod_auth_ldap Information for ftp://rt.fm//pub/OpenBSD/snapshots/packages/amd64/mod_auth_ldap-1.6.0p5.tgz On Tue, May 12, 2009 at 9:07 AM, frantisek holop min...@obiit.org wrote: hi there, for any given package pkg_info starts its info like this, e.g.: amaaq$ pkg_info iodbc-3.52.4p2 Information for inst:iodbc-3.52.4p2 snip my question is, what does that inst: prefix mean, and can it be something else in different situations? -f -- to refuse to decide is a decision.
NEW: devel/libvstr, with a question about lib symlinks
looks like a pretty sane piece of software; it has an extensive test suite, which passes cleanly in my tests on amd64, arm, i386 and sparc64. I had one query, they do the linux-style library symlinks, libvstr.so.[version] - libvstr-1.0.so.[version]. my inclination is just to remove it as in the attached tarball, do you agree? == Vstr is a string library designed for network communication. Its design uses blocks of ptr+length data, so adding, substituting, and deleting data are all fast operations. It has a full API of all the usual string tasks: searching, comparing, splitting, substitution, converting between upper and lower case, and parsing numbers and strings. Has a full POSIX and ISO 9899:1999 compliant printf() implementation including gcc warning compatible custom formatters. == libvstr.tgz Description: application/tar-gz
Re: NEW: devel/libvstr, with a question about lib symlinks
On Tue, 12 May 2009, Stuart Henderson wrote: looks like a pretty sane piece of software; it has an extensive test suite, which passes cleanly in my tests on amd64, arm, i386 and sparc64. I had one query, they do the linux-style library symlinks, libvstr.so.[version] - libvstr-1.0.so.[version]. my inclination is just to remove it as in the attached tarball, do you agree? I though this was acceptable (but don't trust me). IIRC what we definitely don't want is: libfoo.so - libfoo.so.1.0 libfoo.so.1 - libfoo.so.1.0 -- Antoine
special package names
hi there, during upgrade, library packages that are not needed afterwards are renamed to .lib something or something. it is nice, because it shows up right at the top of pkg_info. i think it would make sense to rename all the partial packages as .partial something as well no? as it is, it is interesting to find a bunch of partial packages around pango :] -f -- my favorite mythical creature? the honest politician.
Re: special package names
On Tue, May 12, 2009 at 04:22:33PM +0200, frantisek holop wrote: hi there, Hi, during upgrade, library packages that are not needed afterwards are renamed to .lib something or something. these partial packages often are still needed. They only consist of the shared libs of the older package. Some packages may still depend on these. You could try to remove them with # pkg_delete /var/db/pkg/.libs* it is nice, because it shows up right at the top of pkg_info. i think it would make sense to rename all the partial packages as .partial something as well no? as it is, it is interesting to find a bunch of partial packages around pango :] There may also be partial packages left if the installation of a package is aborted and some files already got installed. These packages are named partial-$packagename. Regards, Markus
PATCH: net/py-boto and sysutils/duplicity
Bruno suggested I take maintainer for these ports. Attached patches change maintainer and update to current versions. Patches mostly from Bruno with minor improvements by me. Duplicity update includes numerous bug fixes and reverts to GPL v2. Tested on two i386 servers running nightly backups to S3 for the last few days. Questions || Commit. Jim Index: Makefile === RCS file: /cvs/ports/sysutils/duplicity/Makefile,v retrieving revision 1.7 diff -u -N -p Makefile --- Makefile28 Mar 2009 18:12:50 - 1.7 +++ Makefile12 May 2009 14:29:23 - @@ -2,18 +2,17 @@ COMMENT = encrypted backup using rsync algorithm -MODPY_EGG_VERSION = 0.5.08 +MODPY_EGG_VERSION = 0.5.17 DISTNAME = duplicity-${MODPY_EGG_VERSION} -PKGNAME = ${DISTNAME}p0 CATEGORIES = sysutils HOMEPAGE = http://www.nongnu.org/duplicity/ MASTER_SITES = ${MASTER_SITE_SAVANNAH:=duplicity/} -MAINTAINER = Bruno Bigras br...@burnbox.net +MAINTAINER = Jim Razmus II j...@bonetruck.org -# GPLv3 +# GPLv2 PERMIT_PACKAGE_CDROM = Yes PERMIT_PACKAGE_FTP = Yes PERMIT_DISTFILES_CDROM =Yes Index: distinfo === RCS file: /cvs/ports/sysutils/duplicity/distinfo,v retrieving revision 1.4 diff -u -N -p distinfo --- distinfo9 Feb 2009 00:28:33 - 1.4 +++ distinfo12 May 2009 14:29:23 - @@ -1,5 +1,5 @@ -MD5 (duplicity-0.5.08.tar.gz) = hNo1+gDrR0Hlq1ut5WxAzw== -RMD160 (duplicity-0.5.08.tar.gz) = sZQBobCXRD1QwhFRoiaoYdSEskE= -SHA1 (duplicity-0.5.08.tar.gz) = 2t5dqZ30JM48eEiiWnxfxnukA8o= -SHA256 (duplicity-0.5.08.tar.gz) = CQO3CUInFQV4Lq/7VvSMnO8loK1AaPqubl+hXvaMeYg= -SIZE (duplicity-0.5.08.tar.gz) = 181262 +MD5 (duplicity-0.5.17.tar.gz) = ow+QgWQBvDtc0UbmCxAkgQ== +RMD160 (duplicity-0.5.17.tar.gz) = ctHj0wC4jsVIdsD7fxdwdhTvDKI= +SHA1 (duplicity-0.5.17.tar.gz) = WB2k94PkvkfUiszBaGIxqHwmFLE= +SHA256 (duplicity-0.5.17.tar.gz) = w/A+K3lbSQn4hIAgLG91oiE9E5b5CieT4sYv+jiiECA= +SIZE (duplicity-0.5.17.tar.gz) = 210353 Index: pkg/PLIST === RCS file: /cvs/ports/sysutils/duplicity/pkg/PLIST,v retrieving revision 1.3 diff -u -N -p pkg/PLIST --- pkg/PLIST 9 Feb 2009 00:28:33 - 1.3 +++ pkg/PLIST 12 May 2009 14:29:24 - @@ -3,6 +3,8 @@ bin/duplicity bin/rdiffdir lib/python${MODPY_VERSION}/site-packages/duplicity/ lib/python${MODPY_VERSION}/site-packages/duplicity-${MODPY_EGG_VERSION}-py${MODPY_VERSION}.egg-info +lib/python${MODPY_VERSION}/site-packages/duplicity/GnuPGInterface.py +lib/python${MODPY_VERSION}/site-packages/duplicity/GnuPGInterface.pyc lib/python${MODPY_VERSION}/site-packages/duplicity/__init__.py lib/python${MODPY_VERSION}/site-packages/duplicity/__init__.pyc lib/python${MODPY_VERSION}/site-packages/duplicity/_librsync.so Index: Makefile === RCS file: /cvs/ports/net/py-boto/Makefile,v retrieving revision 1.5 diff -u -N -p Makefile --- Makefile24 Mar 2009 18:14:10 - 1.5 +++ Makefile12 May 2009 14:28:27 - @@ -2,9 +2,9 @@ COMMENT = Python interface to Amazon Web Services -MODPY_EGG_VERSION = 1.6b +MODPY_EGG_VERSION = 1.7a DISTNAME = boto-${MODPY_EGG_VERSION} -PKGNAME = py-${DISTNAME}p0 +PKGNAME = py-${DISTNAME} CATEGORIES = net HOMEPAGE = http://code.google.com/p/boto/ @@ -12,7 +12,7 @@ MASTER_SITES =http://boto.googlecode.com/files/ \ http://pypi.python.org/packages/source/b/boto/ \ http://release.ingeniweb.com/pypi.python.org-mirror/boto/ -MAINTAINER = Bruno Bigras br...@burnbox.net +MAINTAINER = Jim Razmus II j...@bonetruck.org # MIT PERMIT_PACKAGE_CDROM = Yes Index: distinfo === RCS file: /cvs/ports/net/py-boto/distinfo,v retrieving revision 1.4 diff -u -N -p distinfo --- distinfo9 Feb 2009 00:28:01 - 1.4 +++ distinfo12 May 2009 14:28:27 - @@ -1,5 +1,5 @@ -MD5 (boto-1.6b.tar.gz) = 2jXORJ7QvnSj5dn/9Y+dCA== -RMD160 (boto-1.6b.tar.gz) = xMSpxJfhoUdqiV12BS0Mbzqxp04= -SHA1 (boto-1.6b.tar.gz) = FFYCHZMEBSv2zAHWiLxmvaWEmcM= -SHA256 (boto-1.6b.tar.gz) = yAsbi0gxP2wGOwf/WhiisAx+mNCkTu1cCJL1k+vSOPA= -SIZE (boto-1.6b.tar.gz) = 134305 +MD5 (boto-1.7a.tar.gz) = 5k6ZX/IxNFLnjvN93etEwQ== +RMD160 (boto-1.7a.tar.gz) = CEkl1zYFUMedx+k9e3RrxFJyBgo= +SHA1 (boto-1.7a.tar.gz) = RTQopOt0SkWThPpzKYHs1MMHVOE= +SHA256 (boto-1.7a.tar.gz) = of+WfYAe+TuwVmxP0fI29mLzRYXxfUbCSDZZfmwYAd4= +SIZE (boto-1.7a.tar.gz) = 156683 Index: pkg/PLIST === RCS file: /cvs/ports/net/py-boto/pkg/PLIST,v retrieving revision 1.3 diff -u -N -p pkg/PLIST --- pkg/PLIST 9 Feb 2009 00:28:01 - 1.3 +++ pkg/PLIST 12 May 2009 14:28:27 - @@ -28,6 +28,8 @@ lib/python${MODPY_VERSION}/site-packages/boto/ec2/__in
Having trouble with the www/seamonkey port
I have a new OpenBSD-4.5 system. I've got an updated copy of the cvs archive, so I'm sure I have a *very* recent copy of the ports. I wanted a couple of ports installed before I began my main work on this new system, and that was (1) cvsup (this worked fine) and (2) a browser. I picked www/seamonkey. Among other things, the seamonkey port seems to be *requiring* autoconf-2.62. Apparently, which all of the other autoconf ports installed just fine, autoconf-2.62 won't build, it has problems with a macro on line 1641 of lib/m4sugar/m4sugar.m4. I don't understand those complicated macros too well, I can't troubleshoot it. I tried modifying the autoconf port so that it didn't really want the 2.62 version, but it doesn't matter, www/seamonkey is adamant in needing autoconf-2.62. I read into the seamonkey Makefile, it seems to set AUTOCONF_VERSION to 2.13, and I set that into my environment, but it doesn't matter, seamonkey is still requiring version 2.62. I even tried cleaning out the first build of seamonkey (dind't finish, but I was curious if some environment had stayed around somehow) but it made no difference, it STILL wants autoconf-2.62. I have very little experience at this point with OpenBSD ports, so I don't know what else to do. I did check the m4sugar.m4 files from other autoconf versions, but it seems that the m4sugar.m4 file has been radically changed, I can't just borrow some lines from (say) autoconf-2.61, it won't work, the file is too radically modified. I really need some advice. I sure hope I've guessed correctly on what list to ask help from.
Re: Having trouble with the www/seamonkey port
Sounds like you're trying to mix current ports with a stable base: http://www.openbsd.org/faq/faq15.html#PkgFAQ -Nick Chuck Robey wrote: I have a new OpenBSD-4.5 system. I've got an updated copy of the cvs archive, so I'm sure I have a *very* recent copy of the ports. I wanted a couple of ports installed before I began my main work on this new system, and that was (1) cvsup (this worked fine) and (2) a browser. I picked www/seamonkey. Among other things, the seamonkey port seems to be *requiring* autoconf-2.62. Apparently, which all of the other autoconf ports installed just fine, autoconf-2.62 won't build, it has problems with a macro on line 1641 of lib/m4sugar/m4sugar.m4. I don't understand those complicated macros too well, I can't troubleshoot it. I tried modifying the autoconf port so that it didn't really want the 2.62 version, but it doesn't matter, www/seamonkey is adamant in needing autoconf-2.62. I read into the seamonkey Makefile, it seems to set AUTOCONF_VERSION to 2.13, and I set that into my environment, but it doesn't matter, seamonkey is still requiring version 2.62. I even tried cleaning out the first build of seamonkey (dind't finish, but I was curious if some environment had stayed around somehow) but it made no difference, it STILL wants autoconf-2.62. I have very little experience at this point with OpenBSD ports, so I don't know what else to do. I did check the m4sugar.m4 files from other autoconf versions, but it seems that the m4sugar.m4 file has been radically changed, I can't just borrow some lines from (say) autoconf-2.61, it won't work, the file is too radically modified. I really need some advice. I sure hope I've guessed correctly on what list to ask help from.
Re: Having trouble with the www/seamonkey port
Nick Templeton wrote: Sounds like you're trying to mix current ports with a stable base: http://www.openbsd.org/faq/faq15.html#PkgFAQ Actually, Nik, I have a 4 day old install of the latest release, 4.5, and I haven't yet rebuilt it. Unless I misunderstand your release setup, this means I'm NOT running stable, right? OK, I looked over that URL. Looks like I was in error anyhow ... I can't look at the ports checkout until I reboot out of this (can't yet send/receive mail on the new system yet, so I don't yet know if ports has a release tag, or if maybe I have to build a current system to get things to work. For the record, I'm NOT actually seeing any all kinds of crazy errors, just that one port won't build. -Nick Chuck Robey wrote: I have a new OpenBSD-4.5 system. I've got an updated copy of the cvs archive, so I'm sure I have a *very* recent copy of the ports. I wanted a couple of ports installed before I began my main work on this new system, and that was (1) cvsup (this worked fine) and (2) a browser. I picked www/seamonkey. Among other things, the seamonkey port seems to be *requiring* autoconf-2.62. Apparently, which all of the other autoconf ports installed just fine, autoconf-2.62 won't build, it has problems with a macro on line 1641 of lib/m4sugar/m4sugar.m4. I don't understand those complicated macros too well, I can't troubleshoot it. I tried modifying the autoconf port so that it didn't really want the 2.62 version, but it doesn't matter, www/seamonkey is adamant in needing autoconf-2.62. I read into the seamonkey Makefile, it seems to set AUTOCONF_VERSION to 2.13, and I set that into my environment, but it doesn't matter, seamonkey is still requiring version 2.62. I even tried cleaning out the first build of seamonkey (dind't finish, but I was curious if some environment had stayed around somehow) but it made no difference, it STILL wants autoconf-2.62. I have very little experience at this point with OpenBSD ports, so I don't know what else to do. I did check the m4sugar.m4 files from other autoconf versions, but it seems that the m4sugar.m4 file has been radically changed, I can't just borrow some lines from (say) autoconf-2.61, it won't work, the file is too radically modified. I really need some advice. I sure hope I've guessed correctly on what list to ask help from.
Re: Having trouble with the www/seamonkey port
On 2009/05/12 12:53, Chuck Robey wrote: I have a new OpenBSD-4.5 system. I've got an updated copy of the cvs archive, so I'm sure I have a *very* recent copy of the ports. I wanted a couple of ports installed before I began my main work on this new system, and that was (1) cvsup (this worked fine) and (2) a browser. I picked www/seamonkey. there is only one set of ports which is truly valid for a 4.5 release and that is the set tagged OPENBSD_4_5. please check that is what you have. i would take a guess from the use of cvsup, and building from ports, that you may be more used to FreeBSD; note that here, there is no advantage to building from ports rather than using packages *unless* you need to use a FLAVOR that isn't built by default in the bulk builds, or you need to use a port where license restrictions mean we can't provide packages. when you build from ports on OpenBSD you are building a package *first*, exactly the same as the binary packages we provide (see in /usr/ports/packages after make package), and then afterwards make install just runs pkg_add. I tried modifying the autoconf port so that it didn't really want the 2.62 version, but it doesn't matter, www/seamonkey is adamant in needing autoconf-2.62. i think you're going to have to post some build logs and show what you're doing here, this seems very strange indeed.
Re: NEW: boost-1.39.0
=== Checking files for boost-1.39.0p0 boost_1_39_0.tar.gz doesn't seem to exist on this system. Fetch http://easynews.dl.sourceforge.net/sourceforge/boost/boost_1_39_0.tar.gz. boost_1_39_0.tar.gz 100% |***| 35508 KB00:59 Size does not match for /usr/ports/distfiles/boost_1_39_0.tar.gz I manually changed distinfo but no love there either. Size matches for /usr/ports/distfiles/boost_1_39_0.tar.gz (SHA256) boost_1_39_0.tar.gz: OK === boost-1.39.0p0 depends on: python-2.5* - found === Verifying specs: z c z c === found z.4.1 c.50.1 === Extracting for boost-1.39.0p0 === Patching for boost-1.39.0p0 4 out of 4 hunks failed--saving rejects to Jamroot.rej *** patch-Jamroot did not apply cleanly 1 out of 1 hunks failed--saving rejects to boost/asio/detail/pipe_select_interrupter.hpp.rej *** patch-boost_asio_detail_pipe_select_interrupter_hpp did not apply cleanly 1 out of 1 hunks failed--saving rejects to boost/asio/detail/socket_ops.hpp.rej *** patch-boost_asio_detail_socket_ops_hpp did not apply cleanly 2 out of 2 hunks failed--saving rejects to boost/config/stdlib/libstdcpp3.hpp.rej *** patch-boost_config_stdlib_libstdcpp3_hpp did not apply cleanly 1 out of 1 hunks failed--saving rejects to boost/config/suffix.hpp.rej *** patch-boost_config_suffix_hpp did not apply cleanly 1 out of 1 hunks failed--saving rejects to boost/regex/config.hpp.rej *** patch-boost_regex_config_hpp did not apply cleanly 1 out of 1 hunks failed--saving rejects to boost/smart_ptr/detail/atomic_count.hpp.rej *** patch-boost_smart_ptr_detail_atomic_count_hpp did not apply cleanly 1 out of 1 hunks failed--saving rejects to libs/config/test/boost_has_clock_gettime.ipp.rej *** patch-libs_config_test_boost_has_clock_gettime_ipp did not apply cleanly 1 out of 1 hunks failed--saving rejects to libs/config/test/boost_has_pthread_yield.ipp.rej *** patch-libs_config_test_boost_has_pthread_yield_ipp did not apply cleanly 2 out of 2 hunks failed--saving rejects to libs/random/random_device.cpp.rej *** patch-libs_random_random_device_cpp did not apply cleanly 1 out of 1 hunks failed--saving rejects to libs/regex/src/wc_regex_traits.cpp.rej *** patch-libs_regex_src_wc_regex_traits_cpp did not apply cleanly 4 out of 4 hunks failed--saving rejects to tools/build/v2/tools/gcc.jam.rej *** patch-tools_build_v2_tools_gcc_jam did not apply cleanly *** Error code 1 Stop in /usr/ports/devel/boost (line 2074 of /usr/ports/infrastructure/mk/bsd.port.mk). *** Error code 1 Stop in /usr/ports/devel/boost (line 1427 of /usr/ports/infrastructure/mk/bsd.port.mk). *** Error code 1 Stop in /usr/ports/devel/boost (line 1967 of /usr/ports/infrastructure/mk/bsd.port.mk). On Sun, May 10, 2009 at 12:37:57PM +0400, Andrej Elizarov wrote: here is port for testing some info for maintainer. --- added Markus's patches for asio patches from 1.34: Jamfile_v2 (Jamroot now) - applied libstdcpp3 - applied suffix - applied regex_config - applied atomic_count (smart_ptr_detail_atomic now) - applied function_* - already in code random - applied regex_v4 - already in code config_test - already in code libs_regex_src - applied libs_regex_test - already in code serialization - needed to exclude wserialization. not applied tools_gcc - applied tools_python - not applied stuff: - boost_graph builded with GraphML - boost_test fail build, need some digg around. so disabled in port. - --without-icu - --without-mpi
Re: boost
Stuart Henderson wrote: On 2009-05-09, Markus Hennecke markus-henne...@markus-hennecke.de wrote: Andrej Elizarov wrote: Lucky day, I need boost for wt. It seems wt depend on boost at least v1.35. Markus, do you need any help with porting wt? Sure, I could use some help. If you use the asio port, then boost 1.34.1 will be enough I think. I Was this a separate port just to avoid touching the main boost port, or is there another reason? No other reason. asio was included into boost from 1.35 AFAIR, but updating boost is much more complicated than copying the asio headers into /usr/local or let a package do that. Kind regards, Markus
Re: boost
The headers should have been copied into the boost directory. I am currently fighting this :-( On Tue, May 12, 2009 at 10:20:14PM +0200, Markus Hennecke wrote: Stuart Henderson wrote: On 2009-05-09, Markus Hennecke markus-henne...@markus-hennecke.de wrote: Andrej Elizarov wrote: Lucky day, I need boost for wt. It seems wt depend on boost at least v1.35. Markus, do you need any help with porting wt? Sure, I could use some help. If you use the asio port, then boost 1.34.1 will be enough I think. I Was this a separate port just to avoid touching the main boost port, or is there another reason? No other reason. asio was included into boost from 1.35 AFAIR, but updating boost is much more complicated than copying the asio headers into /usr/local or let a package do that. Kind regards, Markus
x11/eterm,kanji: fix package name
If FULLNAME-${FLAVOR} ever was supported, it certainly hasn't been for a long time now. Index: Makefile === RCS file: /cvs/ports/x11/eterm/Makefile,v retrieving revision 1.26 diff -u -p -r1.26 Makefile --- Makefile8 May 2009 10:25:37 - 1.26 +++ Makefile12 May 2009 20:32:07 - @@ -6,8 +6,7 @@ COMMENT-kanji= color japanese terminal e V= 0.9.4 DISTNAME= Eterm-$V DISTFILES= ${DISTNAME}${EXTRACT_SUFX} -PKGNAME= eterm-$Vp0 -FULLPKGNAME-kanji= ja-eterm-$Vp1 +PKGNAME= eterm-$Vp0 # also see below SHARED_LIBS += Eterm1.0 # .0.0 CATEGORIES=x11 @@ -46,6 +45,7 @@ FLAVORS= kanji FLAVOR?= .if ${FLAVOR:L:Mkanji} +FULLPKGNAME= ja-eterm-$V CATEGORIES+= japanese CONFIGURE_ARGS+=--enable-multi-charset=kanji MODULES+= converters/libiconv -- Christian naddy Weisgerber na...@mips.inka.de
Re: boost
So asio by itself doesn't fix my problem; I need the integrated version with boost. On Sat, May 09, 2009 at 12:13:07PM +0200, Markus Hennecke wrote: Andrej Elizarov wrote: Lucky day, I need boost for wt. It seems wt depend on boost at least v1.35. Markus, do you need any help with porting wt? Sure, I could use some help. If you use the asio port, then boost 1.34.1 will be enough I think. I attached the port, it has both libwthttp and libwtfcgi. libwthttp seems to be working fine, I did some test programs with it and everything looked all right, but I did not get the fastcgi part running. The test programs compile but they fail somehow if run from the apache in base. This could be either some error in wt or some stupid configuration error on my side. This was the first time for me using fastcgi. Kind regards, Markus
python now requires X11 installed
Hello, I used to run all my OpenBSD servers with no X11. Now, to build nmap, you seem to need python. There are no FLAVORS and so no obvious way to prevent it from requiring python. And to build python, you need X11R6 installed. There are no FLAVORS as there used to be, to compile without tkinter to avoid requiring X11R6. Is it possible to install these without installing X11R6? Maybe I just don't understand the way ports work well enough. And, why the change? -- Obama Nation | My emails do not have attachments; it's a digital signature that your mail program doesn't understand. | http://www.subpacefield.org/~travis/ If you are a spammer, please email j...@subspacefield.org to get blacklisted. pgpZmAC7Pyq2D.pgp Description: PGP signature
Re: python now requires X11 installed
I used to run all my OpenBSD servers with no X11. Now, to build nmap, you seem to need python. There are no FLAVORS and so no obvious way to prevent it from requiring python. And to build python, you need X11R6 installed. There are no FLAVORS as there used to be, to compile without tkinter to avoid requiring X11R6. Is it possible to install these without installing X11R6? Maybe I just don't understand the way ports work well enough. And, why the change? Would you prefer to live in a world with 500 flavors? Perhaps. Would you prefer if the ports developers spent all their time building and testing flavors, instead of bringing new code into the ports tree? That is the choice. So sometimes changes like that come direct from the upstream. Supporting flavors (which are mostly restrictions on what gets used) always requires more testing. Sometimes there is no choice but to accept the upstream.
Re: python now requires X11 installed
On Tue, 12 May 2009, travis+ml-pyt...@subspacefield.org wrote: Hello, I used to run all my OpenBSD servers with no X11. Now, to build nmap, you seem to need python. There are no FLAVORS and so no obvious way to prevent it from requiring python. And to build python, you need X11R6 installed. There are no FLAVORS as there used to be, to compile without tkinter to avoid requiring X11R6. Is it possible to install these without installing X11R6? Maybe I just don't understand the way ports work well enough. Why don't you use packages? -- Antoine
Re: python now requires X11 installed
On 2009/05/12 16:01, travis+ml-pyt...@subspacefield.org wrote: Hello, I used to run all my OpenBSD servers with no X11. Now, to build nmap, you seem to need python. There are no FLAVORS and so no obvious way to prevent it from requiring python. And to build python, you need X11R6 installed. There are no FLAVORS as there used to be, to compile without tkinter to avoid requiring X11R6. Is it possible to install these without installing X11R6? Maybe I just don't understand the way ports work well enough. And, why the change? leaving aside the question of what nmap would be doing on a server, pkg_add nmap needs neither python nor X.
Re: python now requires X11 installed
On Tue, May 12, 2009 at 04:01:26PM -0500, travis+ml-pyt...@subspacefield.org wrote: Now, to build nmap, you seem to need python. There are no FLAVORS and so no obvious way to prevent it from requiring python. building ports without X11 installed isn't supported, just use packages (the nmap package doesn't require X11).
Re: python now requires X11 installed
On Tue, May 12, 2009 at 11:10:46PM +0200, Antoine Jacoutot wrote: Why don't you use packages? I just tried, and that solved my problem. I tried to install python from there and it worked just fine. I liked the ports system but understand Theo's points. I have this script which basically does a lot of post-installation work, installing a lot of ports, doing some manual configuration, etc. Now, I'm not quite sure how to convert it to use packages. The problem is that it was relatively easy to cd /usr/ports/foo/bar make install clean. If I use the packages system, I seem to have to know what version I want to install. This seems like it'd be really unstable, since the versions will change from release to release. So to convert the script to work with any release, I'd have to figure out a way to get a listing and select the latest release from the listing. Is there an easy way to do this? -- Obama Nation | My emails do not have attachments; it's a digital signature that your mail program doesn't understand. | http://www.subpacefield.org/~travis/ If you are a spammer, please email j...@subspacefield.org to get blacklisted. pgpbHqnaofZk4.pgp Description: PGP signature
Re: python now requires X11 installed
At 04:20 PM 5/12/2009 -0500, travis+ml-pyt...@subspacefield.org wrote: On Tue, May 12, 2009 at 11:10:46PM +0200, Antoine Jacoutot wrote: Why don't you use packages? Is there an easy way to do this? Try your own variation of: export PKG_PATH=ftp://your favorite mirror/pub/OpenBSD/`uname -a | cut -d -f 3`/packages/`uname -a | cut -d -f 5`/ Identifies the repository automatically by kernel version. Lee
Re: python now requires X11 installed
On Tue, 12 May 2009, travis+ml-pyt...@subspacefield.org wrote: On Tue, May 12, 2009 at 11:10:46PM +0200, Antoine Jacoutot wrote: Why don't you use packages? I just tried, and that solved my problem. I tried to install python from there and it worked just fine. I liked the ports system but understand Theo's points. I have this script which basically does a lot of post-installation work, installing a lot of ports, doing some manual configuration, etc. Now, I'm not quite sure how to convert it to use packages. The problem is that it was relatively easy to cd /usr/ports/foo/bar make install clean. If I use the packages system, I seem to have to know what version Nope, that is not needed. I want to install. This seems like it'd be really unstable, since the versions will change from release to release. So to convert the script to work with any release, I'd have to figure out a way to get a listing and select the latest release from the listing. Is there an easy way to do this? export PKG_PATH=http://path/to/mirror/pub/OpenBSD/`uname -r`/packages/`arch -s/ pkg_add nmap squid openldap foobar ... -- Antoine
Re: special package names
hmm, on Tue, May 12, 2009 at 04:51:31PM +0200, Markus Lude said that during upgrade, library packages that are not needed afterwards are renamed to .lib something or something. these partial packages often are still needed. They only consist of the shared libs of the older package. Some packages may still depend on these. You could try to remove them with i think it would make sense to rename all the partial packages as .partial something as well no? as it is, it is interesting to find a bunch of partial packages around pango :] There may also be partial packages left if the installation of a package is aborted and some files already got installed. These packages are named partial-$packagename. yes, thanks for filling in the details. my suggestion was to make the partial-$packagename show up as .partial-$packagename so it's right at the top with the other special packages, needed or not... -f -- if r is reverse, how come d is forward?
Re: python now requires X11 installed
Thanks all, package stemming drastically simplified my auto-installer. It was much more fragile before, since, say, /usr/ports/mail/mutt/snapshot might or might not be there in new releases. -- Obama Nation | My emails do not have attachments; it's a digital signature that your mail program doesn't understand. | http://www.subpacefield.org/~travis/ If you are a spammer, please email j...@subspacefield.org to get blacklisted. pgpBPdutjidga.pgp Description: PGP signature
Re: python now requires X11 installed
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Matthias Kilian wrote: building ports without X11 installed isn't supported, A useful low-cost, modern solution is a dedicated virtual machine for building ports, which can then be served via http as packages to other systems. That's been working well for me, giving me complete flexibility while avoiding the need for all those messy support libraries on live systems. If you use snapshots or copies, then rolling back the build host to its original clean state after a complex build is trivial. Lather, rinse, repeat. Cheers -d - -- David Talkington dt...@drizzle.com - -- PGP key: http://www.flyingjoke.org/keys/801E3976.asc (What's this? http://en.wikipedia.org/wiki/Digital_signature) -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (OpenBSD) iQEcBAEBAgAGBQJKChpgAAoJEO7jL1CAHjl2QtgH/R8rkkeWeWjWHVG5fplL/udN KKsBBZ/1hjbbS7ZV/uvtkvqkIJFjWaJv48gItA6W/UaMTfKuULTiYc8qtkmVjAXe iDooNsq4aDIVdMK9FZNlF+fSbZDmvE1G/frockAGNY+vkDJV827JGW7qc2D9PZut mWmEkFhZkcpt7gl4ywVCUZywgqSkWhUt2D0fAGCLnZNXk1bsfaZEqb49K0yPsh6I EuPeX3jUdwp5YK0kY93998sLH/PZdKhqRQEKhtXJA9WZI9CHCUzrAqANYDMhpcH5 Y2Iov/S0uaYmBSw7NsJgzhoYy0Dx+aN6EaibCYrmMagSZ0N867r7BnKIhbY+COw= =ghQT -END PGP SIGNATURE-
Re: Having trouble with the www/seamonkey port
Stuart Henderson wrote: On 2009/05/12 12:53, Chuck Robey wrote: I have a new OpenBSD-4.5 system. I've got an updated copy of the cvs archive, so I'm sure I have a *very* recent copy of the ports. I wanted a couple of ports installed before I began my main work on this new system, and that was (1) cvsup (this worked fine) and (2) a browser. I picked www/seamonkey. there is only one set of ports which is truly valid for a 4.5 release and that is the set tagged OPENBSD_4_5. please check that is what you have. i would take a guess from the use of cvsup, and building from ports, that you may be more used to FreeBSD; note that here, there is no advantage to building from ports rather than using packages *unless* you need to use a FLAVOR that isn't built by default in the bulk builds, or you need to use a port where license restrictions mean we can't provide packages. when you build from ports on OpenBSD you are building a package *first*, exactly the same as the binary packages we provide (see in /usr/ports/packages after make package), and then afterwards make install just runs pkg_add. I tried modifying the autoconf port so that it didn't really want the 2.62 version, but it doesn't matter, www/seamonkey is adamant in needing autoconf-2.62. i think you're going to have to post some build logs and show what you're doing here, this seems very strange indeed. Well, couple of comments. First, if you were assuming that I have been using mostly packages in FreeBSD, that's wrong. I won't try to say I've never done it, but it's relatively rare for me to use a package, on either OS. The difference I see, in building your own ports, is that if you insert an option here or there (whether an official one of not), you get exactly whatever you're willing to support yourself. However, since I AM really new at using OpenBSD, I'm going to rebuild my system from -current, so that I can't get bitten by some change in bsd.port.mk. There are some other reasons I have, too, but not worth discussing here. I have spotted and fixed a couple of problems I've already hit... I really need more experience at this system. Please, *don't* take that as ANY port of criticism whatever, it certainly wasn't meant that way. I posted separately about a problem I've hit in src/lib/csu/i386/Makefile. I gotta get my filesystems transformed, quickly, so I can access to/from OpenBSD/FreeBSD, this endless rebooting is getting very old fast. I got a really fine suggestion of using that Linux ext3fs, I need to reread it, but it sure sounds like a fine compromise filesystem.